Are karaoke \k timings used?

No. Karaoke tags live inside braces and are removed with the other override tags. Each line is spoken across its Start to End window; syllable timing is not reproduced.

Do styles, colours or fonts change anything?

No. The [V4+ Styles] section is not read. Two lines in different styles are spoken in the same voice.

Which lines are read aloud?

Every Dialogue line with text left after the tags are stripped, in any style. Comment lines are skipped. Delete sign and on-screen-text lines in the editor before generating.

What about lines that overlap in time?

Both are kept as separate cues and sorted by start time. Review overlaps in the editor and keep the line you want spoken.

Can I upload a classic .ssa file?

The parser reads the same [Events] layout, but uploads are accepted by extension. Rename the file from .ssa to .ass and upload it.

What if my subtitles are SRT or VTT?

Use the SRT-to-audio or VTT-to-audio workflow. All three formats land in the same editor and produce the same kind of timed audio.

Advanced SubStation Alpha · .ass · .ssa

Give your ASS subtitles a voice.
Keep the timing you set.

Upload the .ass file as it left Aegisub. Recastr reads the Dialogue lines, drops the typesetting, and generates a voiceover where every line is spoken in the window you timed it to.

10 free minutes · No credit card required

How it works

1
Upload the .ass file
Straight from Aegisub or a release. No export to SRT.
2
Check which lines will be spoken
Every Dialogue line becomes a cue. Delete sign translations and on-screen text in the editor so only speech is voiced.
3
Pick a voice
One voice for the script, or translate it first and choose a voice for that language.
4
Download MP3 or WAV
Each line is generated into its own start-to-end slot and assembled in time order.

What Recastr reads from an ASS file

An ASS file is a small document with styles, fonts and effects. Only the [Events] section matters for audio. Download a sample .ass that exercises every case below.

[Script Info], [V4+ Styles], fonts and graphics: Ignored. Resolution, wrap style, colours and fonts describe rendering, and nothing is rendered.
The Format line: Read and respected. Field order is taken from the file, so an export with a non-standard column order still parses.
Dialogue and Comment lines: Dialogue lines become cues. Comment lines are skipped, so a line you commented out in Aegisub stays silent.
Override tags: Everything in braces is removed: \pos, \an, \move, \fad, \fs, colour tags and \k karaoke timing. The words are kept.
Line breaks: \N, \n and the hard space \h become ordinary spaces, so a two-line subtitle is one spoken sentence.
Start and End times: H:MM:SS.cc, to the centisecond, converted to milliseconds. Start and End are the only timing used; karaoke syllable timing is not.
Order and layers: Lines are sorted by start time regardless of file order, and layers are flattened. Two lines that overlap on screen both become cues.

Who turns ASS files into audio

Fansub and fandub groups

The translation is already timed line by line. Generate a spoken track from it, or hand the audio to a dubbing team as a timing reference.

Aegisub editors

Keep working in Aegisub. Upload the .ass when the timing is right and let the voice follow it. No re-timing, no SRT export.

Lyric and karaoke videos

Karaoke syllable timing is stripped, but the line timing stays. A spoken translation lands on the same beats as the lyrics.

Language learners

Hear a translated subtitle track spoken over a film or lecture, timed to the original dialogue, without a dubbing studio.

This is not a subtitle renderer

Tools that burn in or preview ASS subtitles care about every style, position and fade. Recastr cares about two things per line: the words and the time window. Everything typographic is discarded on the way in, which is exactly why it never leaks into the voice.

Aegisub files parse as-is

Recastr reads the Format line to find Start, End and Text, so files from Aegisub, Subtitle Edit or a hand-edited release parse without cleanup. Comment lines and sections it does not need are passed over.

Signs and typesetting stay out of the narration

Every Dialogue line is a candidate for speech, including sign translations and title cards. The editor shows them before anything is generated, so remove them there and only dialogue gets a voice.

Watch it work

Frequently asked questions

Are karaoke \k timings used?: No. Karaoke tags live inside braces and are removed with the other override tags. Each line is spoken across its Start to End window; syllable timing is not reproduced.
Do styles, colours or fonts change anything?: No. The [V4+ Styles] section is not read. Two lines in different styles are spoken in the same voice.
Which lines are read aloud?: Every Dialogue line with text left after the tags are stripped, in any style. Comment lines are skipped. Delete sign and on-screen-text lines in the editor before generating.
What about lines that overlap in time?: Both are kept as separate cues and sorted by start time. Review overlaps in the editor and keep the line you want spoken.
Can I upload a classic .ssa file?: The parser reads the same [Events] layout, but uploads are accepted by extension. Rename the file from .ssa to .ass and upload it.
What if my subtitles are SRT or VTT?: Use the SRT-to-audio or VTT-to-audio workflow. All three formats land in the same editor and produce the same kind of timed audio.

Ready to hear your subtitles?

10 free minutes included · No credit card required

Give your ASS subtitles a voice. Keep the timing you set.